Noisy news and exchange rates: A SVAR approach

This paper introduces noisy news shocks into a model of exchange rate determination to measure the impact of these shocks using a SVAR. Agents in the foreign exchange market make decisions with imperfect information about economic fundamentals driving interest rate differentials between countries in that they must rely on a noisy signal of future interest rates. I apply the framework to the USD/GBP nominal exchange rate for the period 1986-2013. Results show that noisy-news explains approximately one fifth of the forecast error variance in the nominal exchange rate, with noise accounting for double (12%) that of news (6%). A historical decomposition of the exchange rate indicates that noise shocks are especially important during periods of changing monetary policy, e.g. the 1990 easing and 2001 tightening of U.S. monetary policy and the unconventional monetary policies surrounding the financial crisis of 2008.